EVI1 dysregulation: impact on biology and therapy of myeloid malignancies - Nature
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Birdwell et al. Blood Cancer Journal (2021)11:64
https://doi.org/10.1038/s41408-021-00457-9 Blood Cancer Journal
REVIEW ARTICLE Open Access
EVI1 dysregulation: impact on biology and therapy
of myeloid malignancies
Christine Birdwell1, Warren Fiskus 1
, Tapan M. Kadia 1
, Courtney D. DiNardo1, Christopher P. Mill1 and
Kapil N. Bhalla 1
Abstract
Ecotropic viral integration site 1 (Evi1) was discovered in 1988 as a common site of ecotropic viral integration resulting
in myeloid malignancies in mice. EVI1 is an oncogenic zinc-finger transcription factor whose overexpression
contributes to disease progression and an aggressive phenotype, correlating with poor clinical outcome in myeloid
malignancies. Despite progress in understanding the biology of EVI1 dysregulation, significant improvements in
therapeutic outcome remain elusive. Here, we highlight advances in understanding EVI1 biology and discuss how this
new knowledge informs development of novel therapeutic interventions. EVI1 is overexpression is correlated with
poor outcome in some epithelial cancers. However, the focus of this review is the genetic lesions, biology, and current
therapeutics of myeloid malignancies overexpressing EVI1.
MECOM locus discovery EVI1: domain-structure and function
1234567890():,;
1234567890():,;
1234567890():,;
1234567890():,;
Evi1 was discovered by Mucenski et al. as a common Human EVI1 is a 145 kilo Dalton (kDa) protein that
site of ecotropic viral integration in mice that caused contains 1051 amino acids. EVI1 localizes to the nucleus
virally induced myeloid malignancies1. Through its rear- and binds DNA through its zinc finger (ZF) domains5.
rangements in human acute myeloid leukemia (AML), the EVI1 contains ten zinc fingers that are arranged in two
human EVI1 gene was mapped to the long arm of chro- separate sets, one N-terminal containing seven zinc fin-
mosome 3 at q 26.2 (3q26.2)2. The EVI1 gene in humans gers, another C-terminal containing three zinc fingers5
is ~92% homologous to the mouse Evi12. EVI1 is encoded (Fig. 1). Through electrophoretic mobility shift assays and
from the MDS1 and ecotropic viral integration site 1 chromatin immunoprecipitation (ChIP) assays, the N-
(EVI1) complex locus (MECOM), which includes several terminal ZF domain was determined to bind TAGA/
alternative transcripts3. EVI1 exists either as a shorter TCTA or CAGAGA/TCTCTG GATA-like simple
single gene or as spliced to the short myelodysplastic sequence repeats (SSRs)5,6. The C-terminal ZF domain
syndrome 1 (MDS1) gene, present more than 350 kb recognizes a CCATATAA ETS-like motif5,6. In the region
upstream to EVI1, creating the longer MDS1-EVI1 gene3. between the ZF domains, is the repressor region that
The shorter isoform of EVI1 is abundant and onco- contains the interaction sites for the co-repressor CtBP
genic4,5. A truncated variant of the EVI1 transcript con- (C-terminal binding protein 1)7,8. EVI1 also contains an
served in both mice and humans, EVI1Δ324, lacks part of acidic domain at its C terminus5 (Fig. 1).
the first zinc finger domain and the ability to transform Disruption of the full-length Evi1 transcript by muta-
(Fig. 1)4. genesis in mice led to severe developmental defects in the
heart and central nervous system, and homozygous
mutants died at approximately embryonic day 10.59.
Additionally, adult mice with conditional knockout of
Evi1 had a marked reduction in their long-term hema-
Correspondence: Kapil N. Bhalla (kbhalla@mdanderson.org)
1
Division of Cancer Medicine, Department of Leukemia, The University of Texas topoietic stem cells (LT-HSCs), and upon transfer into
M. D. Anderson Cancer Center, Houston, TX 77030, USA
© The Author(s) 2021
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction
in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if
changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If
material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain
permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
Blood Cancer JournalBirdwell et al. Blood Cancer Journal (2021)11:64 Page 2 of 14
EVI1 Zinc Fingers (1-7) Zinc Fingers (8-10)
(145 kDa)
Repressive Acidic
1051 aa NH2- Domain Domain -COOH
1 239 547 733 812 886 937
MDS1-EVI1 Zinc Fingers (1-7) Zinc Fingers (8-10)
(200 kDa)
PR Domain Repressive Acidic -COOH
NH2- (188 aa) Domain Domain
PRDF1-RIZ (PR)
homology domain
Truncated Zinc
Finger domain
Zinc Fingers (8-10)
EVI1∆324 1 2 3 4 5
(105 kDa)
Repressive Acidic
NH2- Domain -COOH
Domain
Fig. 1 Schematic of the MDS and EVI1 (MECOM) locus proteins. C2H2-zinc finger motifs are shown in blue. Zinc finger motifs 1–7 and motifs
8–10 form the N-terminal and C-terminal zinc finger domains respectively. The repressive domain and acidic domain are depicted in tan and red,
respectively. The numbers beneath the schematic indicate the amino acid positions of the zinc finger domains, the repressive domain and the acidic
domain of EVI1.
irradiated mice were unable to engraft and repopulate (GATA binding protein 1) and PU.1 (transcription factor
efficiently10. The self-renewal ability of LT-HSCs is linked PU.1)17,19. In the megakaryocyte lineage, EVI1 is expres-
to EVI1 expression, and many LT-HSC-associated genes sed in early precursor cells20. In a transgenic mouse model
have EVI1 binding sites in their regulatory regions11. recapitulating human inv3(q21q26) AML that over-
Furthermore, increased EVI1 expression is a common expresses EVI1 and also has GATA2 haploinsufficiency,
immortalizing factor of murine primary bone marrow EVI1 and GATA2 dysregulation together skewed hema-
after retroviral infection12. For example, MSCV integra- topoiesis toward the megakaryocyte lineage more so than
tion promoted increased expression of EVI1 causing EVI1 overexpression alone21. This suggests that EVI1 may
immortalization of immature myeloid cells, but they were work in concert with other factors to promote the
unable to induce leukemia in transplanted hosts12. Thus megakaryocyte lineage21. In general, for most myeloid
EVI1 supports HSC self-renewal, but EVI1 expression lineages, EVI1 functions to promote a stem or early
alone is not enough to drive leukemogenesis12,13. progenitor transcriptional program11,14. Forced EVI1
In addition to LT-HSC self-renewal, expression of EVI1 expression maintains the stem-like program while
blocks hematopoietic differentiation of the granulocyte, simultaneously suppressing myeloid transcription factors
erythroid, dendritic, and monocytic lineages14,15. EVI1 involved in myeloid differentiation16,18,19,21. Notably,
expression in primary mouse myeloid progenitor cells endogenous EVI1 is generally downregulated under nor-
upregulated HSC-associated genes and decreased DNA mal differentiation14,15. However, the degree to which
replication and repair genes14. EVI1 transcripts are endogenous EVI1 blocks differentiation and what factors
decreased in human CD34+ cells after stimulation of normally downregulate EVI1 during differentiation largely
differentiation induced by cytokine administration, sug- remain an unknown.
gesting that downregulation of EVI1 is an important step
in terminal differentiation of many hematopoietic linea- MDS1-EVI1 and EVI1Δ324
ges15. Forced expression of Evi1 in the mouse bone In 1994, Nucifora et al. identified a transcript of
marrow cell line 32Dcl3 inhibits differentiation response unknown function that they termed MDS1, which formed
to granulocytes and erythrocytes due to granulocyte a fusion protein with RUNX1 and/or EVI1 in several
colony-stimulating factor (G-CSF) and erythropoietin, myelodysplastic syndrome (MDS) patients22. Currently,
respectively16,17. Extrinsic EVI1 expression blocked G- the function of the MDS1 protein itself is still unknown.
CSF-induced differentiation through transcriptional In MDS1-EVI1, exon 2 of MDS1 is fused in-frame to EVI1
repression of the lineage-specific gene myeloperoxidase exon 2, which adds 188 amino acids upstream of the
and the myeloid transcription factors C/EBPα (CCAAT normal start codon of EVI1 in exon 322. A part of these
enhancer binding protein alpha) and RUNX1 (runt-rela- extra N-terminal amino acids contains the PR domain,
ted transcription factor 1, also known as AML1)16,18. which shares homology with the B cell factor positive
Erythroid differentiation was blocked by EVI1 through regulatory domain 1-binding factor (PRD1-BF1) and
binding and subsequent inhibition of transcriptional retinoblastoma binding protein RIZ13,12. The PR domain
activity of the myeloid transcription factors GATA1 is related to a subset of the methyltransferase SET
Blood Cancer JournalBirdwell et al. Blood Cancer Journal (2021)11:64 Page 3 of 14
318 bp Minimal Promoter for EVI1
GATA1 MYB GATA1 ELK1 MYB RUNX1 PU.1 EVI1
GATA2
(TTCTCCTCCTT Transcription
(GGCGATGT) (CACCGTTTCT) (CTGCTTATCTACGT) (GCGATTTCC) (CGAAACGG) (TGCGGTC)
CCCCTCCCTC) Factor Motifs
Fig. 2 Schematic representation of the 318-bp minimal promoter of EVI1 in humans. Within the upstream EVI1 promoter, the positions and
nucleotide sequences/binding sites of specific transcription factors are shown.
domains3,23 (Fig. 1). The combination of a PR domain and binding motifs for RUNX1, ELK1 (ETS transcription
zinc finger domains in MDS1-EVI1 makes it a part of the factor ELK1), RELA (RELA proto-oncogene, NF-κB
PRDM (PR/SET domain) family and thus is also called Subunit), GATA1, and MYB (MYB proto-oncogene,
PRDM3, which was characterized as a mono-methyl transcription factor). Knockdown of RUNX1 and/or ELK1
H3K9 methyltransferase23. Although the specific role of in HEL cells decreased EVI1 mRNA and protein levels26.
MDS1-EVI1 is not always separated from the role of EVI1, Furthermore, interactions between RUNX1 and EVI1 at
loss of MDS1-EVI1 is also associated with embryonic the minimal promoter appear to positively regulate EVI1
lethality, developmental defects, and dysregulation of activity26. MDS1-EVI1 and EVI1Δ324 bind further
hematopoiesis10,15. downstream of the minimal promoter of EVI1 and reduce
EVI1Δ324 is a variant transcript of EVI1 with an its transcription26,27. MDS1-EVI1 and EVI1Δ324 are
internal 972 nucleotide deletion that removes the 6th and reported to be co-expressed with EVI115,24. Although not
7th zinc finger from the N-terminal ZF domain4 (Fig. 1). further studied, present upstream of the minimal EVI1
ChIP assays with FLAG-tagged EVI1 or EVI1Δ324 in an promoter are the consensus binding motifs for GATA1,
ovarian carcinoma cell line (SKOV3) showed an ~71% GATA2, and C/EBPα, suggesting that MDS1-EVI1 and
overlap in binding peaks between the two24. Additionally, EVI1Δ324 may work in concert with other transcription
the transcriptional profile of HeLa cells overexpressing factors to repress EVI127. CML (chronic myeloid leuke-
EVI1-FLAG or EVI1Δ324-FLAG was almost identical24. mia) blast crisis patient-derived cells express high EVI1
However, EVI1Δ324 does not replicate the transformative and β-catenin levels28. Knockdown of β-catenin or its
effects of EVI1 in rat fibroblasts, and is not known to have related co-transcription factor LEF1 (lymphoid enhancer
oncogenic activity nor is it linked currently with any factor 1) decreased EVI1 levels28. Bioinformatic analysis
myeloid malignancy24. indicated two potential tandem LEF1/β-catenin-binding
sites present 1.44 kb upstream of EVI1, which are bound
EVI1 regulation by LEF1, as determined by ChIP assays28. Additional
Epigenetic regulation of EVI1 studies are needed to further clarify regulation of EVI1 by
The region 5′ of EVI1 contains two CpG islands, one LEF1/β-catenin, RUNX1, GATA1, and/or ELK1.
close to the transcription start site of EVI1 and a second
located near MDS125. In an AML cell line that has low Post-translational modifications on EVI1
EVI1 expression, the CpG islands related to EVI1 and EVI1 has been reported to be phosphorylated at serine 196
MDS1 had a marked increase in methylation, suggesting (S196), S538, S858, and S86029,30. Stable isotope labeling of
that EVI1 expression can be regulated by methylation in amino acids followed by mass spectrometry (SILAC-MS)
AML cells25. Furthermore, AML cell lines with high EVI1 identified EVI1-associated proteins. CK2 (casein kinase 2)
expression displayed active chromatin marks, with histone was confirmed to phosphorylate EVI1 residues S538 and
acetylation and enrichment of H3K4me3 (histone 3 lysine S858. Loss of phosphorylation was mediated by PP1α
4 tri-methylation) at the EVI1 promoter. In contrast cell (protein phosphatase 1 alpha), and it decreased DNA-
lines with low EVI1 expression have enrichment of the binding by the C-terminal ZF domain29. In contrast, phos-
repressive histone mark H3K27me3 (histone 3 lysine 27 phorylation of S196 on the 6th zinc finger in the N-terminal
tri-methylation)25. ZF domain decreases DNA binding and repression by EVI1
of promoters containing GATA-like motifs30. Although
EVI1 promoter phosphorylation of Ser858 and Ser860 did not affect EVI1
The minimal promoter of EVI1 was localized to a 318 DNA binding, loss of these phosphorylations blunted EVI1
nucleotide region 5′ of the EVI1 transcription start site transcriptional repression after cellular stress through
that does not contain a traditional TATA or CAAT box26 reduced interaction of EVI1 with co-repressor CtBP131.
(Fig. 2). In the EVI1 minimal promoter, analysis of binding EVI1 is also acetylated by CBP (CREB binding protein or
motifs and site directed mutagenesis identified active KAT3A)/p300 (EP300, or KAT3B) and PCAF (P300/CBP-
Blood Cancer JournalBirdwell et al. Blood Cancer Journal (2021)11:64 Page 4 of 14
associated factor or KAT2B)32. CBP-induced acetylation EVI1 and shown to be required for EVI1 transformation
increased EVI1 transcriptional activity in luciferase assays32. of rat fibroblasts40. This region was also critical for EVI1
In contrast, PCAF-mediated acetylation of EVI1 has been repression of TGF-β (transforming growth factor beta)
reported to exhibit opposing effects on EVI1 activity. Co- signaling and was thus termed the repressive domain (Rp)
expression of EVI1 with PCAF abrogated EVI1-mediated (Fig. 1)41. Two consensus binding motifs for the tran-
Bcl-xL expression, suggesting that EVI1 acetylation blocked scriptional co-repressor CtBP were identified in the EVI1
EVI1 transactivation activity at the Bcl-xL promoter33. In Rp region. The PLDLS sequence at the residue 584 of
contrast, PCAF-mediated acetylation of K564 on EVI1 EVI1 is the major site of CtBP interaction7,8. Mutation of
increased its ability to transactivate GATA2, and this ability the CtBP binding site at residue 584 abolished the ability
was lost in a K564A mutant that cannot be acetylated34. of EVI1 to repress TGF-β-mediated growth arrest and
Overall, it is unclear whether these post-translational mod- transformation of rat fibroblasts7,8.
ifications can occur simultaneously, or whether one mod-
ification can hinder the acquisition of another. Repression of other transcription factors by EVI1
EVI1 can also directly bind several transcription factors
Transcriptional regulation by EVI1 and inhibit their activity (Table 1B). EVI1 was able to
Transcriptional repression by EVI1 repress GATA1-mediated activation of a synthetic pro-
EVI1 co-immunoprecipitates with the H3K9me3 moter. However, EVI1 does not bind to the canonical
methyltransferase SUV39H1 (suppressor of variegation GATA1 motif42. Instead, EVI1 zinc fingers one and six
3–9 homolog 1) and the related H3K9me1/2 methyl- directly interact with the C-terminal zinc finger of
transferase G9a (euchromatic histone lysine methyl- GATA1 in GST-fusion pull-down assays. Also, EVI1
transferase 2)35,36 (Table 1A). EVI1 and SUV39H1 interaction with GATA1 decreased GATA1 DNA-binding
interaction required the N-terminal ZF domain of EVI1 ability. Mutation of EVI1 zinc fingers one and six abol-
and the C-terminal domain of SUV39H1. Histone ished GATA1 interaction and restored differentiation
methyltransferase assays showed SUV39H1 had methyl- potential to 32Dcl3 cells in response to erythropoietin42.
transferase activity alone or in a complex with EVI1. The 6th and 7th zinc finger of EVI1 was shown to
Furthermore, it was observed by the Nucifora and Delwel directly interact with the C-terminal ETS-domain of PU.1
groups that EVI1-mediated repression of a GAL4 luci- through co-immunoprecipitation and GST-fusion pull-
ferase construct was enhanced by SUV39H1 co- down assays. Binding of EVI1 to PU.1 did not prevent
expression35,36. DNA-binding ability of PU.1; instead it blocked associa-
EVI1 represses PTEN (phosphatase and tensin homo- tion of PU.1 with c-Jun (Jun Proto-Oncogene), a subunit
log) through its N-terminal ZF domain and via recruit- of the transcription factor AP-1. Mutation of the 6th and
ment of the polycomb repressor complex 2 (PRC2), 7th EVI1 zinc fingers mitigated EVI1 interaction with
including EZH2 (enhancer of zeste 2), by binding PU.1 and restored differentiation potential to 32Dcl3 cells
upstream of the PTEN transcription start site37. This in response to G-CSF19. The 8th zinc finger in the C-
increased accumulation of the repressive H3K27me3 terminal ZF domain of EVI1 was shown to interact with
mark and reduced histone acetylation at the PTEN locus RUNX143. Binding of EVI1 repressed transcriptional
has been observed in human AML patient samples37. activity of RUNX1 by decreasing its DNA-binding43.
EVI1 interacts through its N-terminal ZF domain with the However, RUNX1 interaction with EVI1 had no effect on
de novo DNA methyltransferases DNMT3A and 3B38,39. EVI1 DNA-binding. EVI1 interacts with the transcription
EVI1 expression correlated with differential hypermethyla- factor SMAD3 through its N-terminal ZF domain41. EVI1
tion of over 200 genes, as compared to normal CD34+ cells, interaction repressed SMAD3 activity leading to blocked
or to a previously reported DNA methylation profile in a TGF-β mediated growth inhibition41.
separate cohort of 344 AML patients39. Unbiased motif
analysis of differentially methylated gene promoters showed Transcriptional activation by EVI1
an enrichment of the motif recognized by the N-terminal A number of gene targets are upregulated by EVI1
ZF domain of EVI139. DNMT3A was also found to be (Table 2). EVI1 interaction with histone acetyltransferases
highly expressed in EVI1-high AML samples compared to has been reported to promote EVI1-mediated transcrip-
other AML subtypes. EVI1 expression levels correlated tional activation32,34. EVI1 interaction with AP-1 subunits
positively with a stronger hypermethylation signature in c-Fos and c-Jun was noticed as early as 1994 by Tanaka
AML patient samples39. et al.44. EVI1-expressing cells exhibited increased c-Fos
and c-Jun levels, and the C-terminal ZF domain of EVI1
Interaction with co-repressor CtBP was critical for activation of the c-Fos promoter44. Loss of
A region just left to the C-terminal ZF domain of EVI1 EVI1 decreased c-Fos occupancy on the DNA, suggesting
was associated with transcriptional repression activity of that EVI1 and AP-1 may act cooperatively at some loci6. A
Blood Cancer JournalBirdwell et al. Blood Cancer Journal (2021)11:64 Page 5 of 14
Table 1 (A) EVI1 interactions with epigenetic regulators. (B) Biology of direct interaction of EVI1 with other
transcription factors.
(A) EVI1 interaction domain Cellular models studied Ref
DNA methyltransferase
DNMT3A N-terminal zinc finger domain 293T, SB1690CB 38,39
DNMT3B N-terminal zinc finger domain 293T, SB1690CB 38,39
Histone methyltransferase
SUV39H1 N-terminal zinc finger domain φE, 293T, HeLa 35,36
G9a N-terminal zinc finger domain φE, 293T, HeLa 35,36
EZH2 N-terminal zinc finger domain THP-1, Jurkat, AML samples 37
Histone acetyltransferase
32
CBP Central region Cos7
32–34
PCAF N-terminal region/C-terminal region Cos7, HT-29, UCSD-AML1
(B) TFs Activity EVI1 interaction domain Cellular models studied Biological outcome Ref
Myeloid
RUNX1 Down 8th zinc finger and central domain NIH-3T3, 32Dcl3, 293T, K562 Blocks myeloid differentiation 43
GATA1 Down 1st and 6th zinc fingers 32DEpo1, 32Dcl3, Cos7, AML14.3D10 Blocks myeloid differentiation 42
PU.1 Down 6th and 7th zinc fingers 32Dcl3 and 293T Blocks myeloid differentiation 19
General
SMAD3 Down 1st–7th zinc fingers 32Dcl3 Blocks TGF-β responsiveness 41
(A) Epigenetic regulator proteins experimentally determined to interact with EVI1.
EVI1 ecotropic viral integration site 1, N-ter ZF domain N-terminal zinc finger domain, DNMT3A/B DNA methyltransferase 3A/B, SUV39H1 suppressor of variegation 3-9
homolog 1, G9a euchromatic histone lysine methyltransferase 2, EZH2 enhancer of zeste 2, CBP CREB binding protein a.k.a. KAT3A, PCAF P300/CBP-associated factor a.
k.a. KAT2B.
(B) Transcription factors experimentally determined to directly interact with EVI1, the interacting domain of EVI1 involved and the implications of the interaction on
the activity of the transcription factor.
TFs transcription factors, EVI1 ecotropic viral integration site 1, RUNX1 RUNX family transcription factor 1, GATA1 GATA binding protein 1, PU.1 transcription factor PU.1,
SMAD3 SMAD family member 3, NF-κB p65 nuclear factor kappa B family member p65.
SILAC-MS screen also confirmed c-Fos and c-Jun inter- abnormalities, molecular alterations, pathological fea-
action with EVI129. This screen also identified several tures, and poor prognosis46,50–52. In inv,(3) breaks most
additional transcription factors and co-factors that inter- frequently occur in a region between RPN1 (Ribophorin
act with EVI1, and 65% of EVI1-regulated genes were 1) and C3orf27, downstream of GATA2, that contains a
upregulated6,29. This highlighted the role of EVI1 as a distal GATA2 hematopoietic enhancer (−77 kb, G2DHE)
transcriptional activator. and the region between C3orf50 and the first exon of the
MECOM locus that encodes for the MDS1-EVI1 tran-
EVI1 dysregulation in myeloid leukemia script (Fig. 3)53. The EVI1 and EVI1Δ324 transcripts
Chromosome 3 lesions leading to EVI1 overexpression remain intact, but the MDS1-EVI1 transcript is frequently
In the World Health Organization (WHO) classification not expressed53. In t(3;3), the breakpoint frequently
of AML and related neoplasms, inversion or translocation occurs in between the MDS1 promoter and the first EVI1
of chromosome 3 at the MECOM locus [inv3(3;3) exon, and MDS1-EVI1 transcript is frequently lost
(q21q26)/inv,(3) t(3;3)(q21;q26.2)/t(3;3)] have been (Fig. 3)46. In 2014, the Delwel group and the Yamamoto
recognized as recurrent genetic abnormalities45 (Fig. 3). group identified that a new super enhancer of ~40 kb is
Inv(3)/t(3;3) is observed in ~1–2.5% of MDS and in a formed from repositioning of the GATA2 distal hema-
similar percentage of AML patients46,47. Inv(3)/t(3;3) topoietic enhancer that drives increased EVI1 expression
rearrangements can also be observed in up to 25–40% of in inv(3)/t(3;3) AML53,54. The new enhancer region gen-
CML patients in blast crisis48,49. Despite their existence as erated by the chromosomal rearrangement was noted to
distinct clinical entities, MDS, AML, and CML with inv contain a 9-kb region with a p300 binding site that
(3)/t(3;3) rearrangements have similar cytogenetic interacts with the EVI1 promoter, and removal of this
Blood Cancer JournalBirdwell et al. Blood Cancer Journal (2021)11:64 Page 6 of 14
Table 2 Transcriptional targets of EVI1.
Gene Activity/levels Regulation Cellular models studied Biological outcome Ref
73
MYC Up Transcriptional upregulation SKOV3, HeLa Active metabolism and apoptosis resistance
33
BcL-xL Up Transcriptional upregulation HT-29, 293T Apoptosis resistance
AML samples
71
GPR56 Up Transcriptional upregulation UCSD-AML1, HNT-34 Apoptosis resistance
AML samples
72
ITGA6 Up Transcriptional upregulation UCSD-AML1, HNT-34 Apoptosis resistance
AML samples
44
c-Fos Up Transcriptional upregulation P19, SKOV3, HeLa Activates AP-1
68
PBX1 Up Transcriptional upregulation HEL Maintains AML stem cell phenotype
Primary murine BM
34
GATA2 Up Transcriptional upregulation EML-C1, HEL Maintains AML stem cell phenotype
Primary mouse EC
17,18
C/EBPα Down Transcriptional repression 32Dcl3, EML, DA-1, U937 Blocks differentiation
16
RUNX1 Down Transcriptional repression 32Dcl3 Blocks differentiation
37
PTEN Down Transcriptional repression Primary murine BM AML Activates metabolism and apoptosis resistance by
samples PI3K/AKT/mTOR pathway
Transcriptional targets of EVI1, the effect on the activity/levels of each target, and the biologic consequence of EVI1-mediated transcriptional regulation on the
target genes.
EVI1 ecotropic viral integration site 1, BM bone marrow, EC embryonic cells, PI3K phosphoinositide 3-kinase, AKT AKT serine/threonine kinase, mTOR mechanistic target
of rapamycin kinase, PTEN phosphatase and tensin homolog, MYC MYC proto-oncogene, BHLH transcription factor, BcL-xL BCL2 Like 1, GPR56 adhesion G protein-
coupled receptor G1, ITGA6 integrin subunit alpha 6, c-Fos Fos proto-oncogene, AP-1 transcription factor subunit, PBX1 PBX homeobox 1, GATA2 GATA binding protein
2, C/EBP CCAAT enhancer binding protein, CDK2 cyclin dependent kinase 2.
binding site attenuates EVI1 expression53. Transgenic interaction domain 1B), respectively55. Similar to the
mice in which the human inv(3) chromosomal abnorm- “enhancer hijacking” of the GATA2 distal hematopoietic
ality was recapitulated through a bacterial artificial chro- enhancer in inv(3)/t(3;3), atypical 3q26 rearrangements
mosome developed leukemia, but if the relocated GATA2 also seem to overexpress EVI1 through repurposing of
enhancer region was deleted, EVI1 expression declined enhancer elements from the translocation partners. In ten
and leukemia did not develop54. These seminal studies cases of atypical 3q26 rearrangements, EVI1 was over-
confirmed that EVI1 dysregulation in response to chro- expressed and the translocation partner whose enhancer
mosomal rearrangements in myeloid disease occurred not was repurposed had decreased expression, with the
from the typical generation of a fusion transcript, but exception of MYC in t(3;8)(q26;q24)55.
rather from “enhancer hijacking”53,54. This highlighted a
two-fold impact, one leading to overexpression of EVI1 EVI1 fusion proteins
and the second causing haploinsufficiency of GATA2, as Several translocations involving the MECOM locus do
it is no longer expressed in the rearranged chromosome53. result in the generation of fusion proteins. The two most
common being t(3;12)(q26;p13) and t(3;21)(q26;q24) that
Atypical 3q26 rearrangements result in ETS variant transcription factor 6 (also TEL)-
In ~0.5–1% of AML and MDS patients, atypical chro- EVI1 and RUNX1-MDS1-EVI1 fusion proteins, respec-
mosome 3 rearrangements occur involving the MECOM tively22,56 (Fig. 4). Both translocations are rare, found in
locus55. Most atypical 3q26 rearrangements have levels of less than 1% of myeloid malignancies56,57. Fluorescent
EVI1 overexpression comparable to inv(3)/t(3;3) cases, in situ hybridization demonstrated that the t(3;12)
similar phenotypic changes, and share a poor prog- breakpoints are between the ETV6 exon 2 and 3 and on
nosis46,55. Atypical 3q26 rearrangements include, but are heterogeneous regions in 3q26, both 3′ and 5′ of MDS1 as
not limited to t(2;3)(q21;q26), t(3;7)(q26;q24), t(3;8)(q26; well as in between MDS1 and EVI156. The resulting
q24), and t(3;6)(q26;q25), which involve THADA translocation fuses the first two exons of ETV6 with the
(THADA armadillo repeat containing), CDK6 (cyclin entire MDS1-EVI1 or EVI1 transcript. Since no known
dependent kinase 6), MYC (V-Myc avian myelocytoma- functional domain of ETV6 is added to EVI1 in the fusion
tosis viral oncogene homolog), and ARID1B (AT-rich protein, it is thought that the oncogenic properties of the
Blood Cancer JournalBirdwell et al. Blood Cancer Journal (2021)11:64 Page 7 of 14
Chromosome 3
p24.3
p24.2
p21.3
p21.2
p14.3
p14.2
q13.1
q13.2
q13.3
q25.1
q26.1
q26.2
q26.3
q11.2
p25
p23
p22
p13
p12
q21
q22
q23
q24
q27
q28
q29
p11
Chr3 Normal cells Chr3
q21 q26
GATA2
EVI1
GATA2
Chromatin ~40 Mb
Looping
GATA2 Distal GATA2
Enhancer RPN1 EVI1
~128 kb
Chr3 inv(3)(q21;q26) Chr3
q21 EVI1 EVI1 EVI1 q26
GATA2
Inv(3) BP
Distal GATA2
GATA2 RPN1 Enhancer EVI1
Chromatin
Looping
Chr3 t(3;3)(q21;q26) Chr3
q21 q26
EVI1 EVI1 EVI1
GATA2
t(3;3) BP
Distal GATA2
GATA2 EVI1 RPN1
Enhancer
Fig. 3 Schematic of the q21-26 locus on chromosome 3 in normal cells and cells with inv(3)(q21q26) or t(3;3)(q21q26). In both inv(3)
(q21q26) or t(3;3)(q21q26), the breakpoints lead to juxtaposition of a region surrounding the distal GATA2 enhancer and the RPN1 gene in 3q21 with
the EVI1 gene in 3q26. Breakpoints occur 3′ of the EVI1 gene in the inv(3)(q21q26) setting, whereas they occur 5′ of the EVI1 gene in the case of t(3;3)
(q21q26). In both types of 3q21q26 rearrangement, the GATA2 enhancer induces EVI1 gene transcription instead of GATA2 expression and thus
promotes leukemogenesis.
Acidic
A EVI1 NH2- 1 ZnF 2-7 Zn F 8- 1 0 Domain
-COOH
MDS1
NH2- Acidic
MDS1-EVI1 1 Z n F 2- 7 ZnF 8-10 -COOH
Domain
PR Domain
B
RUNX1 NH2- Runt Domain TAD -COOH
t(3:21)(q26;q24)
RUNX1- Acidic
MDS1-EVI1 NH2- Runt Domain 1 ZnF 2-7 ZnF 8-10 Domain
-COOH
C
ETV6 NH2- HLH ETS -COOH
t(3:12)(q26;p13)
ETV6-EVI1 NH2- Acidic
1 ZnF 2-7 ZnF 8-10 -COOH
Domain
ETV6- Acidic
NH2- 1 ZnF 2-7 ZnF 8-10 -COOH
MDS1-EVI1 Domain
Fig. 4 EVI1 and EVI1 fusion proteins. Schematic diagram of the EVI1, MDS1, MDS1-EVI1, RUNX1, RUNX1-MDS1-EVI1, ETV6, ETV6-EVI1, and ETV6-
MDS1-EVI1 proteins.
Blood Cancer JournalBirdwell et al. Blood Cancer Journal (2021)11:64 Page 8 of 14
fusion protein come from the inappropriate expression 7 in the dominate clone, suggesting that EVI1 could favor
and function of EVI1 driven by the ETV6 promoter56. In expansion of clones with monosomy 7 or that EVI1 could
line with this, similar to other 3q26 rearrangements, contribute to the genomic instability leading to monos-
myeloid malignancies expressing the ETV6-EVI1 fusion omy 713.
are associated with dysmegakaryopoiesis and poor prog-
nosis58. The t(3;21) can generate RUNX1-MDS1-EVI1 Effect of EVI1 on hematopoietic stem cell proliferation/
fusion protein, where the DNA-binding RUNT domain of differentiation
RUNX1 is fused to the whole MDS1-EVI1 protein (Fig. EVI1 is known to directly interact with and repress the
4)59. Expression of the RUNX1-MDS1-EVI1 protein is activity of a number of myeloid transcription factors
associated with disruption of RUNX1 and EVI1 regulatory including GATA1, PU.1, and RUNX117,19,42,43. Enforced
networks. This is thought to be partly achieved by tran- Evi1 expression transcriptionally repressed C/EBP-α in
scriptional repression of RUNX1 targets through murine hematopoietic cells18. EVI1-mediated repression
recruitment of co-repressors by EVI1 in the fusion pro- of C/EBP-α was also observed in the murine hemato-
tein57. In mouse models of conditional RUNX1-MDS1- poietic progenitor cell line 32Dcl317. Further confirma-
EVI1 expression or transplant models, the fusion protein tion that EVI1 represses C/EBP family members is
is associated with development of hematopoietic dysplasia needed through in vivo leukemia models and in patient-
and acute megakaryoblastic leukemia60. derived samples. EVI1 also regulates hematopoietic dif-
ferentiation and proliferation through transcriptional
EVI1 overexpression without chromosome 3 aberrations repression of several miRNAs. EVI1 repressed miR-9
Aberrant EVI1 expression can also occur in the absence levels through binding to its regulatory region, recruiting
of chromosome 3 rearrangements. EVI1 overexpression is DNMT3B, and inducing DNA methylation66. Decreased
observed in ~8–10% of MDS, 8% of de novo AML, and miR-9 led to increased levels of its target genes FOXO1
30% of advanced CML, but it is unclear here how EVI1 and 3 (Forkhead Box O1 and 3)66. EVI1 expression was
overexpression occurs61. Several ChIP studies have shown also found to decrease miR-449A levels, and ChIP ana-
that mixed lineage leukemia (MLL) and MLL fusion lysis showed EVI1 bound miR-449A regulatory region67.
proteins, including MLL-AF9 and MLL-ENL bind to the Repression of miR-449A by EVI1 increased expression of
EVI1 regulatory region, resulting in increased EVI1 the miR-449A-targets Notch1 and Bcl-2 in human AML
expression62,63. In a recent report in which MLL-AF9 cell lines67.
fusion gene was expressed either in murine Sca−Kit+ EVI1 transcriptionally activates the hematopoietic
(LSK) HSCs or in granulocyte monocyte precursors proto-oncogene PBX1 (PBX homeobox 1) through bind-
(GMPs), LSK-MLL-AF9 cells had significantly higher ing to its promoter region68. Knockdown of PBX1
levels of Evi1 than GMP-MLL-AF9 cells64. Additionally, decreased EVI1-mediated transformation of primary
AMLs with high EVI1 expression have been shown to be mouse bone marrow cells68. Comparing tissues from wild
associated with inferior relapse-free and overall survival65. type to those from EVI1+/− and EVI1−/− mice, at
embryonic day 9.5, GATA2 expression was decreased in
Biologic consequences of 3q lesions and EVI1 EVI1 depleted tissues69. EVI1 expression also correlated
overexpression with high expression of megakaryocytic markers, includ-
Genomic instability ing the thrombopoietin receptor MPL70. Furthermore, in
Utilizing SILAC-MS studies to determine EVI1 inter- a mouse model of EVI1 leukemia, thrombopoietin
action partners, Bard-Chapeau et al. observed enrichment expression correlated with EVI1 expression, and double
in protein domains associated with DNA repair, chro- positive EVI1-thrombopoietin cells had enhanced sec-
matin remodeling, and transcription29. Furthermore, the ondary leukemia formation ability in a serial bone marrow
EVI1 N-terminal ZF domain binds to GATA-like SSRs, transplant assay70. Collectively, in myeloid malignancies
and EVI1 ChIP analysis revealed an increase in recombi- expressing EVI1, transcriptional alterations of specific
nation rates near EVI1 bound SSR13,24. How EVI1 myeloid transcription factors, and of miRNAs, contribute
increases genomic instability is not well characterized to myeloid dysplasia.
beyond its protein interactions. However, a gene therapy
study using a Maloney murine leukemia virus vector to Increased drug resistance
express NADPH-oxidase conducted in two patients to Several pathways have been implicated in EVI1-
treat chronic granulomatous disease unfortunately caused mediated resistance to apoptosis leading to drug-
integration of the vector at the MECOM locus. The resistance. High EVI1 expression correlated with high
patients developed clonal expansion of myeloid cells expression of the anti-apoptotic Bcl-xL protein in CML
bearing activating insertions in the MECOM locus and patient samples33. Conversely, knockdown of EVI1 was
EVI1 overexpression. Both patients developed monosomy shown to decrease Bcl-xL levels by approximately five
Blood Cancer JournalBirdwell et al. Blood Cancer Journal (2021)11:64 Page 9 of 14
Table 3 Retrospective analysis of clinical outcome of patients with 3q26 genetic lesions.
Year First author N CR (%) Median OS (m) 1-year OS Long term OS Long term relapse probability Ref
46
2010 Lugthart 79 31% 10.3 N.D. 5-year OS: 5.7% 5-year RFS: 4.3%
75
2010 Grimwade 69 36% N.D. N.D. 10-year OS: 3% 10-year CIR: 89%
45
2011 Sun 30 42% 8.9 33% 5-year OS: 3% N.D.
87
2015 Wanquet 40 29% 10.6 N.D. 4-year OS: 3% N.D.
76
2020 Sitges, M 61 29% 8.4 42% 4-year OS: 13% N.D.
N number, ORR overall response rate, CR complete remission, OS overall survival, m months, RFS relapse-free survival, CIR cumulative incidence of relapse, N.D. not
discussed.
fold33. EVI1 interactions with the microenvironment are poor response to therapy, and has been linked with
also implicated in apoptosis-resistance. The adhesion acquisition of resistance to tyrosine kinase inhibitors48.
molecules ITGA6 (integrin subunit alpha 6) and GPR56
(adhesion G protein-coupled receptor G1) are highly Monosomy 7 and MLL translocations
expressed in EVI1-positive AML, and their knockdown Loss of one copy of chromosome 7 (monosomy 7, −7)
leads to increased apoptosis in response to Ara-C treat- or deletion of the long arm of chromosome 7 (−7q) is
ment and loss of RhoA (ras homolog family member A) observed in 30–70% of MDS and AML with inv(3)/t
signaling, respectively71,72. In AML, cells high EVI1 (3;3)74. Retrospective studies have shown that inv(3)/t(3;3)
expression correlated with high MYC and BCL2 expres- MDS/AML with −7/−7q display worse prognosis than
sion, with poorer clinical outcome73. inv(3)/t(3;3) alone74,77. Which genetic alteration occurs
first is unclear, and likely varies on a case-by-case basis,
Clinical phenotypes and outcome of EVI1-positive given the heterogeneity of the myeloid malignancies. As
myeloid malignancies noted above, in two cases where gene therapy activated
MDS with EVI1 overexpression is commonly associated EVI1 expression through retroviral insertion, both cases
with dyserythropoiesis and with the presence of micro developed monosomy 7 in the dominant leukemic clone,
megakaryocytes51. Categorized as high risk, more than suggesting that EVI1 at least favors events leading to
half of inv(3)/t(3;3) MDS patients with EVI1 over- monosomy 713. The q arm of chromosome 7 contains
expression progress to AML within ~2 years of diag- several key genes whose haploinsufficiency is considered
nosis46,47. Furthermore, EVI1 overexpression correlates to be a loss of tumor-suppressor and thus contribute to
with shorter overall survival and poorer response to leukemia transformation. These genes include EZH2 and
treatment51. Overall survival of patients with EVI1- MLL3, as well as the cytoplasmic cellular regulators
positive MDS ranges from 13 to 17 months after diag- SAMD9 (Sterile Alpha Motif Domain Containing 9) and
nosis46,47. Like EVI1-positive MDS, AML with EVI1 SAMD9L78. Perhaps due to the ability of MLL and MLL-
overexpression often presents with myeloid dysplasia, fusion proteins to upregulate EVI1 transcription, EVI1
particularly of the erythrocyte and megakaryocytic linea- overexpression can be observed in ~30% of cases with
ges46,51. Studies have also reported EVI1 expression as an MLL translocation, and here EVI1 expression correlates
independent prognostic factor for poorer overall survival with poor prognosis63,65.
in AML, and high EVI1 expression is associated with
poorer response to therapy46,74,75. Several clinical studies Transcription factor mutations
have reported that, in 3q26-rearranged AML, the median Approximately 20% of MDS and AML patients with inv
overall survival after diagnosis remains approximately less (3)/t(3;3) express mutations in RUNX179. Another tran-
than 1 year, whereas long-term overall survival is less than scription factor IKZF1 (IKAROS family zinc finger) is also
15% (Table 3)47,74,76. mutated in up to 25% of cases of inv(3)/t(3;3) MDS or
EVI1-expressing CML may also be associated with AML. Since IKZF1 is located on chromosome 7, IKZF1
megakaryocytic dysplasia51. EVI1 expression is rarely mutations occur in clones without chromosome 7 dele-
detected in the chronic phase of CML, but is readily tions77. Although not a mutation, almost all MDS, AML,
detected in a significant proportion (25–40%) of blast and CML with inv(3)/t(3;3) have GATA2 haploinsuffi-
crisis of CML, suggesting that acquisition of EVI1 ciency due to the re-location of the GATA2 distal
expression can drive progression into blast crisis48,49. hematopoietic enhancer53,54. This was shown to con-
EVI1 expression in CML blast crisis is correlated with tribute to EVI1-driven leukemia transformation21. Of
Blood Cancer JournalBirdwell et al. Blood Cancer Journal (2021)11:64 Page 10 of 14
note, despite loss of expression from one allele of GATA2, or t(3;21), therapy with a second generation tyrosine
15% of inv(3)/t(3;3) can carry additional mutations in kinase inhibitors and/or chemotherapy is utilized. EVI1-
GATA2 on the non-rearranged allele52,77. positive myeloid malignancies have been documented to
be relatively refractory to current therapies. There is no
Activating mutations in signaling pathways statistical difference in the overall 5-year survival rates
A significant proportion of inv(3)/t(3;3) MDS and AML between MDS and AML with inv(3)/t(3;3), which avera-
cases have activating mutations in RAS GTPase family ges at 3–5%46,47,51,74. In a study by the Delwel group,
member (NRAS or KRAS), or in other RAS-signaling allogenic hematopoietic stem cell transplant following the
pathway proteins, including PTPN11 (protein tyrosine first clinical remission yielded increased survival odds in
phosphatase non-receptor type 11), and NF1 (neurofi- AML patients with MLL translocation with EVI1 over-
bromin 1), which promote dysregulated RAS signaling expression65. However, greater than 40% of the patients
and uncontrolled proliferation52,77,79. These mutations are still succumbed to their disease65. Overall survival rate of
observed in 66–98% of inv(3)/t(3;3) MDS/AML52,77,79. A patients with CML who initially respond but later pro-
greater percentage of AML cases with inv(3)/t(3;3) AML gress on TKI therapy and acquire EVI1 overexpression is
carried RAS family mutations, as compared to the MDS half compared to those without EVI1 expression48.
cases79.
Potential targeted therapies for EVI1-positive myeloid
Mutations in epigenetic machinery malignancies
Low frequency of mutations in DNMT3, TET2 (tet To date, following treatment of myeloid malignancies
methylcytosine dioxygenase 2), and IDH1/2 (isocitrate with inv(3)/t(3;3) or EVI1 overexpression with targeted
dehydrogenase 1/2) were observed in AML or MDS with therapies, including DNA hypomethylating drugs, vene-
inv(3)/t(3;3)46,51,77. However, mutations in the polycomb toclax or glasdegib, or with FLT3 TKI or IDH1/2 inhibi-
group protein ASXL1 (ASXL transcriptional regulator 1) tors, clinical outcome data are unavailable80. A targeted
were reported in ~20% of AML cases with inv(3)/t(3;3)77. agent has yet to be identified and developed that exhibits
Mutations in splicing factors SF3B1 (splicing factor 3b clinical efficacy specifically against EVI1-overexpressing
subunit 1) and U2AF1 (U2 small nuclear RNA auxiliary myeloid malignancies.
factor 1) were also found in ~30–60% of inv(3)/t(3;3)
MDS or AML cases52,77. The biologic impact of these Treatment with ‘epimodifiers’
‘epimutations’ in myeloid malignancies on the transcrip- One promising target is the chromatin reader protein
tional signature attributed to inv(3)/t(3;3) and EVI1 BRD4 (bromodomain containing 4), which is involved in
overexpression remains to be elucidated. transcriptional activation, especially via sustaining the
activity of super enhancers, such as those of MYC, CDK4/
Mutations inversely correlated with EVI1 6 and BCL2/Bcl-xL81. By also inhibiting GATA2 super
Mutations in NPM1 (nucleophosmin 1) and C/EBP-α enhancer, treatment with BET (bromodomain and extra-
inversely correlate with inv(3)/t(3;3) and EVI1 expres- terminal motif) inhibitor could repress EVI1, as well as
sion55,77,79. Why EVI1 overexpression is not seen with reverse EVI1-dependent transcriptional programs
NPM1 or C/EBP-α mutations is unknown. One possibility through inactivating enhancers and super enhancers of
could be that survival of the clones with high EVI1 the key oncogenes (Fig. 5). Preclinical use of BET inhibitor
expression in combination with a NPM1 or C/EBP-α treatment was reported to inhibit growth and induce
mutation is impaired. apoptosis of an EVI1-overexpressing AML cell line82.
Since several BET inhibitors are already undergoing
Clinical outcome with standard therapy of clinical evaluation, they represent an attractive therapy
myeloid malignancies with inv(3)/t(3;3) option for myeloid malignancies with inv(3)/t(3;3) and/or
Standard front-line treatment of MDS includes DNA EVI1 overexpression53.
de-methylating agents like azacitidine and decitabine. In Several transcriptional regulators including EVI1 are
advanced, high-risk MDS carrying inv(3)/t(3;3) with an acetylated by CBP/p30032. For example, RUNX1 interacts
increased percent of bone marrow blasts between 5 and with EVI1 and is positively regulated by acetylation32,83,84.
20%, or in overt transformation of MDS to secondary Furthermore, the GATA2 distal enhancer that is relocated
AML (sAML), chemotherapy with cytarabine (Ara-C) and and transactivates EVI1 in inv(3) and t(3;3) chromosomal
the anthracyclines, idarubicin and daunorubicin is com- aberrations contains a p300 binding site that is critical in
monly employed. In CML, treatment in the chronic phase driving EVI1 expression54. Additionally, the GATA2
generally begins with a tyrosine kinase inhibitor (TKI) enhancer has increased read-through of enhancer RNAs
such as imatinib, dasatinib or nilotinib. In transformation (eRNAs) at the breakpoints that cause its repositioning,
of CML into blast crisis (CML-BC) carrying inv(3)/t(3;3) and the CBP/p300 inhibitor GNE-049 is reported to
Blood Cancer JournalBirdwell et al. Blood Cancer Journal (2021)11:64 Page 11 of 14
Targeting EVI1 Transcription and Activity in MDS/AML with BET Inhibitors
anti-apoptotic Bcl-xL suggests that BH3-mimetic inhi-
Hijacked GATA2 bitor targeting this anti-apoptotic protein could also have
enhancer therapeutic value, alone or in combinations against EVI1-
BET expressing myeloid malignancies. Recently, the small
inhibitor
EVI1 promoter molecule compound pyrrole-imidazole polyamide, which
inhibits the DNA-binding activity of N-terminal ZF
domain of EVI1, was shown to induce apoptosis of EVI1-
Immune
expressing AML cells due to downregulation of the EVI1
Metabolism EVI1 surveillance target GRP56, which was linked to EVI1-mediated resis-
c-Myc NF-κB p65 tance to apoptosis71. Several therapies targeting anti-
AKT/mTOR
apoptotic proteins are in clinical trials, and these agents
Stem phenotype
Environmental could potentially be employed in treatment of EVI1-
interaction
Pbx1 GPR56
expressing myeloid malignancies.
Growth/
GATA2 Survival ITGA6
SMAD3
GATA1
RUNX1
c-Myc Targeting β-catenin-TCF7L2 activity
AKT/mTOR
C/EBPα c-Fos In CML with inv(3)/t(3;3), β-catenin/TCF1 (T cell factor
C/EBPε Bcl-xL
GPR56
1) signaling was reported to be activated, which positively
ITGA6 regulated EVI128. Therefore, along with treatment with
SMAD3
BCR-ABL1 targeted TKI inhibition of β-catenin/TCF
Fig. 5 Effects of targeting EVI1 transcription and activity in MDS/ signaling may be a promising strategy. Additionally, sev-
AML with BET inhibitors. In AML cells with inv(3)(q21q26) or t(3;3) eral groups have demonstrated that resistance to BET
(q21q26), the hijacked GATA2 enhancer interacts with the EVI1
promoter resulting in high expression of EVI. This leads to aberrant
inhibitors in AML is mediated by the activity of β-catenin/
regulation of multiple transcriptional programs including metabolism, TCF7L2/c-Myc axis, resulting in re-expression of c-Myc
stem cell phenotype, growth/survival, environmental interactions, and despite treatment with BET inhibitor89. In this setting
immune surveillance in AML cells. Treatment with BET inhibitors that also, treatment with a β-catenin/TCF signaling inhibitor,
evict BET proteins such as BRD4 from the chromatin of the GATA2 e.g., BC2059 (tegavivint), may not only reverse BET
enhancer leads to downregulation of EVI1 and its activity in AML cells.
inhibitor resistance but also exhibit synergy with BET
inhibitor against AML with inv(3)/t(3;3) and/or EVI1
overexpression90.
particularly inhibit eRNAs53,85. Therefore, targeted com-
bination therapy with this HAT inhibitor would simulta- Summary and future directions
neously cause deacetylation of the transcription factors, While great strides have been made, there is still much
leading to decreased transcriptional activity, as well as more to be elucidated regarding EVI1 biology and its
cause the loss of super enhancer function, thereby effec- contributions to leukemogenesis. EVI1 is a transcriptional
tively shutting down EVI1 transcriptional program. regulator that promotes a stem-like expression program in
EVI1 can repress transcription through recruitment of hematopoietic progenitors, crucial to their self-renewal,
DNMT3A or B resulting in de novo methylation39,86. growth, and repopulating potential. The role of co-
EVI1 expression is also associated with hypermethylation expression of alternative EVI1 transcripts in myeloid
of over 200 genes in AML samples39,86. Consistent with malignancies remains to be determined. EVI1 regulates
this, DNA methyltransferase inhibitors have exhibited transcription through recruitment of epigenetic modifiers.
clinical activity in EVI1-overexpressing AML87. However, Recent studies have highlighted the dysregulated tran-
since monotherapy with DNA hypomethylating agent as a scriptome and signaling pathways that underpin the
first-line treatment for MDS exhibits only a modest aberrant biology, aggressive phenotype, and refractoriness
clinical efficacy, combination with other targeted thera- to standard therapy of EVI1-overexpressing myeloid
pies is likely to achieve superior efficacy against EVI1- malignancies (Fig. 5). How commonly the co-occurring
expressing MDS. Use of EZH2 inhibitor to abrogate genetic alterations and mutations, e.g., monosomy 7 and
dependency of EVI1-expressing MDS or AML with RAS pathway mutations, and their order of acquisition,
monosomy 7 on the residual normal EZH2 function may contribute to the aggressive phenotype and therapy-
exert added efficacy. refractoriness of EVI1-overexpressing myeloid malig-
nancies has yet to be fully characterized. New probes or
Targeted therapy with BH3-mimetic apoptosis inducers alternative methodology need to be developed that will
Recently, co-treatment with venetoclax, a BH3-mimetic assist in probing the clonal architecture of the co-
inhibitor of Bcl-2 protein, with azacitidine was approved mutations that occur with EVI1 dysregulation at the
for therapy of AML88. EVI1-mediated upregulation of the single-cell level. Importantly, functional genomic studies
Blood Cancer JournalBirdwell et al. Blood Cancer Journal (2021)11:64 Page 12 of 14
need to be conducted to identify specific dependences that 7. Izutsu, K. et al. The corepressor CtBP interacts with Evi-1 to repress trans-
can be targeted to achieve superior efficacy against EVI1- forming growth factor beta signaling. Blood. 97, 2815–2822 (2001).
8. Palmer, S. et al. Evi-1 transforming and repressor activities are mediated by
overexpressing myeloid malignancies. Large scale screens CtBP co-repressor proteins. J. Biol. Chem. 276, 25834–25840 (2001).
with or without the presence of a promising therapeutic 9. Hoyt, P. R. et al. The Evi1 proto-oncogene is required at midgestation for
agent, e.g., by CRISPR technology, could also potentially neural, heart, and paraxial mesenchyme development. Mech. Dev. 65, 55–70
(1997).
yield new knowledge for designing effective combination 10. Zhang, Y. et al. PR-domain-containing Mds1-Evi1 is critical for long-term
therapies. Analysis of the 3D chromatin architecture of hematopoietic stem cell function. Blood. 118, 3853–3861 (2011).
“hijacked enhancers” at EVI1 locus and their response to 11. Kataoka, K. et al. Evi1 is essential for hematopoietic stem cell self-renewal, and
its expression marks hematopoietic cells with long-term multilineage repo-
treatment may also provide insights into effective ways to pulating activity. J. Exp. Med. 208, 2403–2416 (2011).
decrease EVI1 expression, suggest promising treatments, 12. Du, Y., Jenkins, N. A. & Copeland, N. G. Insertional mutagenesis identifies genes
and potential mechanisms of resistance. How dysregulated that promote the immortalization of primary bone marrow progenitor cells.
Blood. 106, 3932–3939 (2005).
EVI1 expression in myeloid malignancies creates immune 13. Stein, S. et al. Genomic instability and myelodysplasia with monosomy 7
evasion and T cell exhaustion also remains to be fully consequent to EVI1 activation after gene therapy for chronic granulomatous
elucidated. In this exciting era of novel immunotherapies disease. Nat. Med. 16, 198–204 (2010).
14. Kustikova, O. S. et al. Activation of Evi1 inhibits cell cycle progression and
new research avenues and potential strategies have already differentiation of hematopoietic progenitor cells. Leukemia. 27, 1127–1138
been illuminated. These are likely to involve harnessing of (2013).
the innate or adaptive immune mechanisms to overcome 15. Steinleitner, K. et al. EVI1 and MDS1/EVI1 expression during primary human
hematopoietic progenitor cell differentiation into various myeloid lineages.
immune tolerance or T cell exhaustion in eliminating Anticancer Res. 32, 4883–4889 (2012).
myeloid malignancies, including those driven by EVI1 16. Morishita, K., Parganas, E., Matsugi, T. & Ihle, J. N. Expression of the Evi-1 zinc
dysregulation. finger gene in 32Dc13 myeloid cells blocks granulocytic differentiation in
response to granulocyte colony-stimulating factor. Mol. Cell. Biol. 12, 183–189
(1992).
Conflict of interest
17. Kreider, B. L., Orkin, S. H. & Ihle, J. N. Loss of erythropoietin responsiveness in
T.M.K. receives research funding from BMS/Celgene, Amgen, AstraZaneca,
erythroid progenitors due to expression of the Evi-1 myeloid-transforming
Astellas, Pfizer, AbbVie, Genentech, JAZZ, Cellenkos, InCyte, and Ascentage; T.
gene. Proc. Natl Acad. Sci. USA 90, 6454–6458 (1993).
M.K. serves as a consultant/advisory board member for Agios, Pfizer, Abbvie,
18. Wilson, M. et al. EVI1 Interferes with Myeloid Maturation via Transcriptional
Genentech, JAZZ, Daiichi Sankyo and Novartis. C.D.D. receives research funding
Repression of Cebpa, via Binding to Two Far Downstream Regulatory Ele-
from Abbvie, Agios, Calithera, Cleave, BMS/Celgene, Daiichi-Sankyo,
ments. J. Biol. Chem. 291, 13591–13607 (2016).
ImmuneOnc and Loxo. C.D.D. serves as a consultant to Abbvie, Agios, Celgene/
19. Laricchia-Robbio, L., Premanand, K., Rinaldi, C. R. & Nucifora, G. EVI1 Impairs
BMS, Daiichi Sankyo, ImmuneOnc, Novartis, and Takeda. C.D.D. serves on an
myelopoiesis by deregulation of PU.1 function. Cancer Res. 69, 1633–1642
advisory board with stock options for Notable Labs. K.N.B. serves as a
(2009).
consultant to Iterion Therapeutics. All other authors have no conflict of interest
20. Shimizu, S. et al. EVI1 is expressed in megakaryocyte cell lineage and enforced
to declare.
expression of EVI1 in UT-7/GM cells induces megakaryocyte differentiation.
Biochem. Biophys. Res. Commun. 292, 609–616 (2002).
21. Yamaoka, A. et al. EVI1 and GATA2 misexpression induced by inv(3)(q21q26)
Publisher’s note contribute to megakaryocyte-lineage skewing and leukemogenesis. Blood
Springer Nature remains neutral with regard to jurisdictional claims in Adv. 4, 1722–1736 (2020).
published maps and institutional affiliations. 22. Nucifora, G. et al. Consistent intergenic splicing and production of multiple
transcripts between AML1 at 21q22 and unrelated genes at 3q26 in (3;21)(q26;
Received: 7 January 2021 Revised: 25 February 2021 Accepted: 3 March q22) translocations. Proc. Natl Acad. Sci. USA 91, 4004–4008 (1994).
2021 23. Pinheiro, I. et al. Prdm3 and Prdm16 are H3K9me1 methyltransferases required
for mammalian heterochromatin integrity. Cell. 150, 948–960 (2012).
24. Sayadi, A. et al. Functional features of EVI1 and EVI1Delta324 isoforms of
MECOM gene in genome-wide transcription regulation and oncogenicity.
References Oncogene. 35, 2311–2321 (2016).
1. Mucenski, M. L. et al. Identification of a common ecotropic viral integration 25. Vazquez, I. et al. Down-regulation of EVI1 is associated with epigenetic
site, Evi-1, in the DNA of AKXD murine myeloid tumors. Mol. Cell. Biol. 8, alterations and good prognosis in patients with acute myeloid leukemia.
301–308 (1988). Haematologica. 96, 1448–1456 (2011).
2. Morishita, K. et al. The human Evi-1 gene is located on chromosome 3q24-q28 26. Maicas, M. et al. Functional characterization of the promoter region of the
but is not rearranged in three cases of acute nonlymphocytic leukemias human EVI1 gene in acute myeloid leukemia: RUNX1 and ELK1 directly reg-
containing t(3;5)(q25;q34) translocations. Oncogene Res. 5, 221–231 (1990). ulate its transcription. Oncogene. 32, 2069–2078 (2013).
3. Fears, S. et al. Intergenic splicing of MDS1 and EVI1 occurs in normal tissues as 27. Maicas, M. et al. The MDS and EVI1 complex locus (MECOM) isoforms regulate
well as in myeloid leukemia and produces a new member of the PR domain their own transcription and have different roles in the transformation of
family. Proc. Natl Acad. Sci. USA 93, 1642–1647 (1996). hematopoietic stem and progenitor cells. Biochim. Biophys. Acta. Gene Regul.
4. Bordereaux, D., Fichelson, S., Tambourin, P. & Gisselbrecht, S. Alternative spli- Mech. 1860, 721–729 (2017).
cing of the Evi-1 zinc finger gene generates mRNAs which differ by the 28. Manachai, N. et al. Activation of EVI1 transcription by the LEF1/beta-catenin
number of zinc finger motifs. Oncogene. 5, 925–927 (1990). complex with p53-alteration in myeloid blast crisis of chronic myeloid leu-
5. Perkins, A. S., Fishel, R., Jenkins, N. A. & Copeland, N. G. Evi-1, a murine zinc kemia. Biochem. Biophys. Res. Commun. 482, 994–1000 (2017).
finger proto-oncogene, encodes a sequence-specific DNA-binding protein. 29. Bard-Chapeau, E. A. et al. EVI1 oncoprotein interacts with a large and complex
Mol. Cell. Biol. 11, 2665–2674 (1991). network of proteins and integrates signals through protein phosphorylation.
6. Bard-Chapeau, E. A. et al. Ecotopic viral integration site 1 (EVI1) regulates Proc. Natl Acad. Sci. USA 110, E2885–E2894 (2013).
multiple cellular processes important for cancer and is a synergistic partner for 30. White, D. J. et al. Phosphorylation of the leukemic oncoprotein EVI1 on serine
FOS protein in invasive tumors. Proc. Natl Acad. Sci. USA 109, 2168–2173 196 modulates DNA binding, transcriptional repression and transforming
(2012). ability. PLoS ONE 8, e66510 (2013).
Blood Cancer JournalYou can also read
